bell e., harrison t.m., mcculloch m.t. and young e.d. (2011)

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 75 (2011) 4816–4829

Early Archean crustal evolution of the Jack Hills Zirconsource terrane inferred from Lu–Hf, 207Pb/206Pb, and

d18O systematics of Jack Hills zircons

Elizabeth A. Bell a,⇑, T. Mark Harrison a,b, Malcolm T. McCulloch c,d,Edward D. Young a

a Dept. of Earth and Space Sciences, University of California at Los Angeles, Los Angeles, CA 90095, USAb Institute of Geophysics and Planetary Physics, University of California at Los Angeles, Los Angeles, CA 90095, USA

c Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australiad School of Earth and the Environment, University of Western Australia, Crawley, Perth, WA 6009, Australia

Received 1 September 2010; accepted in revised form 26 May 2011; available online 14 June 2011

Abstract

Several lines of isotopic evidence – the most direct of which is from Hadean Jack Hills zircons – suggest a very early historyof crust formation on Earth that began by about 4.5 Ga. To constrain both the fate of the reservoir for this crust and thenature of crustal evolution in the sediment source region of the Jack Hills, Western Australia, during the early Archean,we report here initial 176Hf/177Hf ratios and d18O systematics for <4 Ga Jack Hills zircons. In contrast to the significant num-ber of Hadean zircons which contain highly unradiogenic 176Hf/177Hf requiring a near-zero Lu/Hf reservoir to have separatedfrom the Earth’s mantle by 4.5 Ga, Jack Hills zircons younger than ca. 3.6 Ga are more radiogenic than –13e (CHUR) at3.4 Ga in contrast to projected values at 3.4 Ga of –20e for the unradiogenic Hadean reservoir indicating that some later juve-nile addition to the crust is required to explain the more radiogenic younger zircons. The shift in the Lu–Hf systematicstogether with a narrow range of mostly mantle-like d18O values among the <3.6 Ga zircons (in contrast to the spread towardssedimentary d18O among Hadean samples) suggests a period of transition between 3.6 and 4 Ga in which the magmatic settingof zircon formation changed and the highly unradiogenic low Lu/Hf Hadean crust ceased to be available for intracrustalreworking. Constraining the nature of this transition provides important insights into the processes of crustal reworkingand recycling of the Earth’s Hadean crust as well as early Archean crustal evolution.� 2011 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

The suggestion that the silicate Earth differentiated toform a continental-like crust during its first few hundredmillion years was, until recently, highly controversial. Incontrast to the traditional paradigm of continental growthoccurring largely since 4 Ga or later (e.g., Taylor andMcLennan, 1985), isotopic evidence has recently emerged

0016-7037/$ - see front matter � 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2011.06.007

⇑ Corresponding author. Address: 595 Charles E. Young Dr. E,3806 Geology Building, Los Angeles, CA 90095, USA. Tel.: +1 310206 2940.

E-mail address: [email protected] (E.A. Bell).

suggesting that enriched, possibly continental reservoirsdeveloped substantially before that time during the so-called Hadean eon. For example, evidence for very early(>4.35–4.53 Ga) differentiation of the silicate earth has beeninferred from 142Nd/144Nd variations in terrestrial samples(Caro et al., 2003) and the contrast between 142Nd/144Nd inchondrites and the silicate Earth (Boyet and Carlson, 2005),although this is also explicable in terms of a non-chondriticbulk silicate Earth (Dauphas and Chaussidon, 2011). Ndisotopic evidence for a region of enriched Hadean crust isinferred from 142Nd/144Nd data for amphibolites from theNuvvuagittuq Greenstone Belt (O’Neill et al., 2008), butinconsistent with other data (e.g., Cates and Mojzsis, 2009).

http://dx.doi.org/10.1016/j.gca.2011.06.007

mailto:[email protected]

http://dx.doi.org/016/j.gca.2011.06.007

Granitoids, ~2.7 Ga

Dugel Gneiss, ~3.4 Ga

Meeberrie Gneiss, ~3.5-3.7 Ga

Metasediments, Mid-Archean

100 km

Mt. Narryer

Jack Hills

YilgarnCraton

26˚S

27˚S

116˚E 117˚E

Fig. 1. Geologic sketch map of the Narryer Gneiss Complex, afterMyers (1988a). Location of Jack Hills and Mt. Narryer indicated.

Early Archean Jack Hills zircon Lu–Hf 4817

Independent evidence for early felsic crust comes fromthe 176Hf/177Hf compositions of detrital Jack Hills zircons,Narryer Gneiss Complex, Western Australia (Harrisonet al., 2005, 2008; Blichert-Toft and Albarede, 2008). Zir-cons tend toward low Lu/Hf ratios, such that the ingrowthof 176Hf from beta decay of 176Lu is typically minimal. Thuszircons reflect, with minimal correction for radiogenic in-growth, the initial 176Hf/177Hf ratio of their host rock. Asignificant number of Hadean Jack Hills zircons containhighly unradiogenic hafnium, suggestive of derivation froma near-zero Lu/Hf reservoir formed almost immediately fol-lowing accretion of the planet (Harrison et al., 2005, 2008).In addition, the zircons record oxygen isotope and trace ele-ment signatures interpreted to imply the existence of surfacewater (e.g., Mojzsis et al., 2001) and water-saturated gra-nitic melting conditions (Watson and Harrison, 2005) by4.3 Ga, which are also suggestive of continental crust.

One interesting aspect of the Nd and Hf isotopic evi-dence for early crust formation is that the inferred early en-riched reservoir(s) has apparently not been significantlyreworked and sampled by younger rocks. Also, only verysmall quantities of Hadean materials bearing this signaturesurvive. It is likely that this early enriched crust has beendestroyed well before the present day, but examination ofthe <4 Ga portion of the Jack Hills detrital zircon popula-tion should shed light on the extent and longevity of thisreservoir in the Jack Hills source terrane(s) during the earlyArchean.

Although the vast majority of Lu–Hf isotopic analyses(Harrison et al., 2005, 2008; Blichert-Toft and Albarede,2008) have concentrated on the >4 Ga Jack Hills zircons,the detrital population ranges in age from �3 to nearly4.4 Ga (Holden et al., 2009). In this paper, we present thelargest dataset to date (130 analyses) of Lu–Hf measure-ments of <4 Ga Jack Hills zircons and use this record toinvestigate how late into Earth’s history the early low Lu/Hf reservoir is evident in the Jack Hills detrital population.We also evaluate the relative importance of crustal growthvs. reworking over the ca. 1 billion year detrital zircon Lu–Hf isotope record in an effort to constrain the chemical andtectonic evolution of the Yilgarn Craton.

2. GEOLOGIC SETTING

The Jack Hills are located in the northwestern corner ofthe Yilgarn Craton, Western Australia, within the NarryerGneiss Complex (Fig. 1). The Narryer complex consists ofEarly to Late Archean orthogneisses and several metasedi-mentary associations, some of which host >4 Ga zircons.The relationship between the metasediments and orthog-neisses in the Jack Hills and throughout the Narryer terraneis uncertain. The present contacts between the crystallineand metasedimentary units are thought to be tectonicrather than depositional (Nutman et al., 1991; Spaggiariet al., 2007). Despite broad agreements between the agesof younger detrital zircons in Jack Hills metasedimentsand the ages of Narryer gneisses (Nutman et al., 1991;Maas and McCulloch, 1992), whole-rock REE geochemis-try (Maas and McCulloch, 1992) of the metasedimentsand more detailed comparison of the age distributions

(Maas and McCulloch, 1992; Amelin, 1998) suggest thatthe presently exposed Narryer gneisses are distinct fromthe source of Archean zircons in the Jack Hills population.It is therefore best to consider the detrital zircon and Narr-yer gneiss record separately for the purpose of constrainingcrustal evolution in the region.

The largest concentrations of Hadean grains are foundwithin apparently fluvial (Williams and Myers, 1987) peb-ble metaconglomerates likely deposited at �3 Ga (Spagg-iari et al., 2007). U–Pb age surveys of the population tendto reveal maxima in the age distribution at 3.4 and4.1 Ga, the younger being much more prominent, and agap (or minimum) in the distribution between about 3.6and 3.8 Ga (Kober et al., 1989; Amelin, 1998; Crowleyet al., 2005; Holden et al., 2009). Automated SHRIMPage analysis of >100,000 zircons has revealed that concor-dant Hadean (>4 Ga) grains make up approximately 5%of the metaconglomerate population (Holden et al., 2009).

Observable magmatic rocks in the Narryer Gneiss Com-plex span ages from 2.6 to 3.73 Ga (Nutman et al., 1991;Spaggiari et al., 2007) (Fig. 1). A widespread unit, the Mee-berrie gneiss, contains protoliths ranging from 3.73 Ga ton-alites to 3.6 Ga granitoids (Nutman et al., 1991) with aprominent monzogranitic unit at 3.68 Ga (Myers, 1988a)(Fig. 1). The Manfred gabbro-anorthosite complex at3.73 Ga is included in several of the younger gneisses(Myers, 1988a,b). The 3.44–3.49 Ga tonalitic Euradagneisses and the �3.38 Ga syeno- to monzogranitic Dugelgneisses occupy the time period most heavily sampled bydetrital zircons (Myers, 1988a; Nutman et al., 1991). Sev-eral granites and pegmatites dating from �3.0 to 3.3 Gaare also found within the Narryer Complex (Bennettet al., 1990). Finally, 2.6–2.7 Ga granites intrude the region

4818 E.A. Bell et al. / Geochimica et Cosmochimica Acta 75 (2011) 4816–4829

concomitantly with widespread metamorphism, faulting,and folding (Myers, 1988a; Nutman et al., 1991).

Previous work on crustal evolution of the NarryerTerrane focused on the Sm–Nd and U–Pb compositionsof exposed orthogneisses. Maas and McCulloch (1992)recalculated TDM from earlier work (DeLaeter et al.,1985) on the Meeberrie, Eurada and Dugel gneisses and3.0–3.3 Ga granitoids. The tonalitic portions of the Mee-berrie appear to be sourced from relatively juvenile materi-als, whereas the monzogranitic younger Meeberrie, Dugel,and 3.3 Ga granitoids appear to be sourced largely fromolder (but <4 Ga) reworked crust with little juvenile input(Maas and McCulloch, 1992). The Eurada gneiss and the3.0 Ga granitoids appear to be formed from the reworkingof distinct sources extracted from the depleted mantle morerecently than the Meeberrie (Maas and McCulloch, 1992).The high apparent source l of the 3.73 Ga Manfred Com-plex may suggest a substantial component of >4 Ga crust(Fletcher et al., 1988).

Despite the unclear relationship between the sources forJack Hills detrital zircons and the known geology of theNarryer Gneiss Complex, it is clear that the isotopic char-acteristics of some Archean Narryer gneisses record evi-dence for crustal evolution since Hadean times. A largedataset of Lu–Hf data for the younger detrital zirconsmay provide a parallel, but more detailed, account of crus-tal evolution in the region for the early and middleArchean.

3. METHODS

We sampled the Jack Hills zircon distribution with thegoal of obtaining a more complete picture of the age, inter-nal textures, and isotope geochemistry among the youngerdetrital population. We selected an epoxy mount (RSES51)from the study of Holden et al. (2009) containing approxi-mately 255 randomly picked zircons from the Jack Hillsdetrital population at the discovery site (Compston andPidgeon, 1986) from the large collection of Jack Hills grainsanalyzed at the Australian National University. The datingprotocol of Holden et al. (2009) employs automated sam-pling of each grain for several seconds to establish an esti-mate of 207Pb/206Pb. Full U–Pb analysis was done only forgrains with apparent 207Pb/206Pb ages >3.95 Ga. Thereforemost grains employed in this study were not precisely da-ted. One hundred and twenty nine grains on mountRSES51 were analyzed for both Lu–Hf isotopes and207Pb/206Pb age, with the only further selection being forgrains with a large enough uncracked surface area to beused for laser ablation Lu–Hf analyses. This approach,we believe, provides an essentially random sample of thatportion of the zircon population large enough for laserablation sampling.

3.1. Imaging for textures

Imaging was mainly accomplished using a scanning elec-tron microscope. A combination of BSE imaging andEDAX were used to identify mineral inclusions (severalof which were further studied by electron microprobe for

stoichiometry) within the zircons. Cathodoluminescence(CL) imaging was used to elucidate internal structuresand zoning. Zircons were sorted into zoning-style categoriesfollowing the suggestions of Corfu et al. (2003). Categoriesincluded (1) oscillatory zoning, (2) core/rim geometry, (3)patchy or irregular zoning, (4) sector zoning, and (5) no(or too faint to be distinguishable) zoning. Inhomogeneouszircons with uncertain patterns were sorted into the“patchy/irregular” category, which became a catchall forambiguous grains. Due to the possibility for zoning pat-terns to be rendered uncertain by grain fragmentation dur-ing sedimentary cycling, the patchy/irregular category ismost likely over-represented.

3.2. Coupled Hf–Pb measurements

We measured Lu–Hf systematics and 207Pb/206Pb agesfor the zircons by laser ablation multicollector inductivelycoupled plasma mass spectrometry (LA-MC-ICPMS),employing ThermoFinnigan NEPTUNE MC-ICPMS andassociated lasers at the Australian National Universityand UCLA. A combination of magnet switching and zoomoptics switching were used to switch between Lu–Hf and Pbisotopic measurements, adapted for the NEPTUNE fromthe procedure of Woodhead et al. (2004). This method givesus the potential to deconvolve the results of the semi contin-uous analyses into definable Hf–Pb domains, increasing theaccuracy of the interpretations. Though we do not samplethe Lu–Hf and Pb mass sets simultaneously as in the workof, e.g., Xie et al. (2008) (who employ a laser ablation lineleading to two separate mass spectrometers), we are able todetermine coherent age-Hf domains by bracketing our Hfanalyses with Pb analyses showing the same 207Pb/206Pbage. One disadvantage of the approach is the lack of infor-mation regarding U–Pb discordance. Despite its limita-tions, however, this combined Hf–Pb approach is moreuseful for correlating Lu–Hf systematics with an applicableage than the more traditional in situ analysis of U–Pb andLu–Hf information on separate volumes of material. De-spite our fairly large laser spot size (80–100 lm) which runsthe risk of overlapping multiple age domains in the hori-zontal direction, our in situ sampling and ability to detectage domains in the vertical direction gives us a significantadvantage also over solution U–Pb and Lu–Hf methods.

We identified separate Hf–Pb age domains on the basisof the 207Pb/206Pb ratio, requiring the presence of a plateauin 207Pb/206Pb ages for at least three Pb counting cycles. Forage domains in the interior of crystals, only Lu–Yb–Hf data(two back-to-back cycles) bracketed by consistent Pb data(three cycles) were considered. Using this method we iden-tified 130 separate Hf–Pb domains among the grains ana-lyzed. Analyses were accomplished during two sessions:Session One took place at ANU in September 2007 and Ses-sion Two at UCLA in April 2009. Background subtractionwas accomplished online, and all further data reduction wasdone offline.

3.2.1. Session One

Fifty-nine grains (totaling 61 Hf–Pb domains) were ana-lyzed using a 193 nm laser with a circular spot 80 lm in


diameter. Ten seconds of counting on the Yb, Lu, Hf massset (i.e., 171, 173–179, 181) alternated with 3 s of countingon the Pb mass set, with 4 s of magnet settling time betweeneach mass set. A total of 130 s counting time was given toeach analysis. See Harrison et al. (2005, 2008) for detailsof the analytical methods.

3.2.2. Session Two

Session Two took place over three days in April of 2009at UCLA. Statistics related to the accuracy of the peakstripping and mass fractionation corrections were calcu-lated on a day-to-day basis. Sixty-six grains (totaling 69Hf–Pb domains) were analyzed by laser ablation using anExcistar ArF excimer laser with a circular spot 100 lm indiameter. Eleven seconds of counting on the Yb, Lu, andHf mass set alternated with 5 s of counting on the Pb massset (204, 206, 207, 208), and the first 2 s of counting on eachmass set were disregarded during data reduction in order toensure a two-second settling time for the magnet. Thisleaves 9 s counting on the Yb, Lu, and Hf masses alternat-ing with 3 s counting on Pb considered for the final analy-sis. A maximum of 160 s (120 used for data) was given toeach analysis. Blanks were run before each analysis in thissession.

1.4671

1.4672

1.4673

1.4674

1.4675

1.4676

1.4677

1.4678

1.4679

1.4680

871/f

H771

fH

a

0.0082

0.0084

0.0086

0.0088

0.0090

0.0092

471/fH

771f

H

In-Sessio

c

0 20 40 60 80 100

Fig. 2. Hafnium isotopic results for two uniform ratios in nature. (A andanalysis sessions. (C and D) 178Hf/177Hf results after mass fractionation coOne of Session Two are uniformly low compared to the accepted value (Tfor 178Hf/177Hf does not appear to correlate with either deviation from tvalues for 176Hf/177Hf.

3.2.3. Interference and mass fractionation correction

Despite its advantages (e.g., lesser destructivity to thesample), sampling of Lu and Hf in zircon by laser ablationrather than in solution precludes chemical removal of iso-baric interferences. Yb, which occurs in zircon at the tracelevel (Finch and Hanchar, 2003), presents interferences withHf at masses 174 and 176. Isotopes of Lu and Hf at mass176, the relevant isotopes involved in the Lu–Hf decay sys-tem, also mutually interfere. Analysis by laser ablation thusnecessitates peak-stripping to deconvolve the signals atmass 174 and 176. Details of the Hf isotopic analysis andpeak stripping procedures for isobaric interference of Yband Lu on Hf isotopes are given in Harrison et al. (2005,2008) and Taylor et al. (2008). We tested the accuracy ofthe peak stripping and fractionation corrections (normal-ized to (179Hf/177Hf = 0.7325) by comparing the corrected174Hf/177Hf and 178Hf/177Hf values for each analysisagainst the accepted values (0.008657 ± 5 and 1.46735 ±16 respectively, ±2r) of Thirlwall and Anczkiewicz (2004)(Fig. 2). Our corrected values for 174Hf/177Hf, which ishighly sensitive to the veracity of the 174Yb stripping, agreewell on average with the reference value (Fig. 2c and d).

In Session 1, four analyses out of 102 (�4%) fell morethan 3r from the reference value. In Session 2, four

b

n Analysis #0 20 40 60 80 100 120 140 160

d

Day 1 Day 2 Day 3

Day 1 Day 2 Day 3

B) 174Hf/177Hf results after mass fractionation correction for bothrrection for both analysis sessions. Results for Session One and Dayhirlwall and Anczkiewicz, 2004). Deviation from the accepted valuehe reference value for 174Hf/177Hf or from the standards’ accepted

Fig. 4. Comparison of results from ANU and UCLA Hf–Pbsessions in eHf vs. age space. One high Lu/Hf measurement at2.88 Ga, �28e was omitted for space purposes.


analyses out of 139 (�3%) fell more than 3r from the ref-erence value: one analysis on Day 1 (N = 34), one analysison Day 2 (N = 50), and two analyses on Day 3 (N = 55).Our corrected values for 178Hf/177Hf are uniformly low by0.85e for Session 1 and by 1.29e for Day 1 of Session 2,which occurred one week before days 2 and 3 of Session2. During the ANU session, nine out of 102 analyses(including both unknowns and standards) fell more than3r from the reference value (Fig. 2a and b). Statistics forthe peak-stripping accuracy for the UCLA session were cal-culated on a day-by-day basis.

This higher than expected incidence of analyses inconsis-tent with standard values for the isotope ratios is balancedby fractionation- and interference-corrected 176Hf/177Hfvalues for the AS3 and Mud Tank standard zircons thatagree well with the reference values of Woodhead and Her-gt (2005) (see Fig. 3). In both sessions, mass-discriminationcorrections were interpolated from the standard values’offsets and applied to the unknowns bracketed by each setof standards. The standards AS3, Temora-2, and Mudtankwere used for calculating 176Hf/177Hf offsets; only AS3 andTemora-2 were used for calculating 176Lu/177Hf owing toMudtank’s very low Lu/Hf ratio compared to our un-knowns (nearly two orders of magnitude below Temora-2and AS3; Woodhead and Hergt, 2005). The results comparewell between the two sessions (see Fig. 4), lending confi-dence to our conclusion that the data reduction and correc-tion procedures have yielded accurate results. In addition tothe standard zircons, the NIST610 standard glass was usedas a secondary 207Pb/206Pb standard. 207Pb/206Pb of stan-dards was sensitive to the 207Pb average signal, becomingunreliable at a threshold of 10 millivolts (Fig. 5). The stan-dards with 207Pb signals above the threshold fell within twopercent of their known values, whereas standards with low-er signals were considerably more inaccurate (Fig. 5). Onlystandards falling above the threshold were considered; allunknowns fell above the threshold as well. Only NIST610and AS3 standards fell above the threshold and were subse-quently used for comparing to unknowns. This result isunsurprising: the Mudtank and Temora-2 zircon standards

a

Fig. 3. 176Hf/177Hf results for the standard zircons AS3 and Mudtank, forOne data were collected in September 2007 at ANU; (b) Session Two da

are much younger than AS3 (732 and 417 Ma compared to1099 Ma, respectively) and they consequently have lowerPb contents. The Mudtank and Temora-2 Pb data reflectthe physical limitations of the technique at low Pb signal– notably lower signal than any of our unknowns or thanthe relatively Pb-rich AS3 and NIST610 standards. Becauseof the good agreement of the high-signal standards withtheir accepted 207Pb/206Pb ages, we did not apply a massfractionation correction to any of our Pb data. We analyzedat mass 204 and 208, and found no appreciable 204Pb in oursamples. 208Pb/206Pb values did not display any trend indic-ative of common Pb contamination (see e.g., Blichert-Toftand Albarede, 2008). Given the lack of noticeable commonPb contamination among our zircons, we did not apply acommon Pb correction to the data.

Fractionation – and interference-corrected 176Lu/177Hffor the standard zircons AS3 and Temora-2 compare wellwith the reference values of Woodhead and Hergt (2005).In-day averages for 176Lu/177Hf, 176Hf/177Hf, and 207Pb/

b

which analytical conditions are closest to the unknowns. (a) Sessionta were collected in April 2009 at UCLA.

1E-3 0.01 0.1

-2

0

2

4

6

8

10

12

14

16

Sess. 1 Pb standards Sess. 2 Pb standards

% D

evia

nce

from

acc

epte

d va

lue

207Pb Total signal (V)

1E-3

0.01

0.1

1

Sess. 2 Unknowns

702Pb

Tot

al S

igna

l (V)

Fig. 5. All Pb standard analyses in % deviance (from expectedvalue) vs. total 207Pb signal (V) space. No standards were omittedfor Session 1, but Session 2 standards below 0.01 V were omitted.Inset: Session 2 unknowns; all are >0.01 V.


206Pb of standard zircons are tabulated together with theaccepted values in Table 1. Uncertainties are based bothon internal uncertainty and the reproducibility of the rele-vant standards for each quantity.

All eHf values are calculated using the CHUR values ofBouvier et al. (2008), and a decay constant for 176Lu of1.867 � 10�11 yr�1 (Scherer et al., 2004; Soderlund et al.,2004; Amelin, 2005). CHUR was likely not a physicalreservoir for long in the early Archean, and so we adopt itonly as a well-defined reference value. eHf uncertainties in-clude uncertainties in 176Hf/177Hf, 176Lu/177Hf, contempora-neous CHUR values of these ratios, and uncertainty in the176Lu decay constant using the error propagation formulaeof Harrison et al. (2008). Lu–Yb–Hf/Pb data for all stan-dards and unknowns are shown in Electronic annex EA-1.

3.3. Oxygen isotopes

Analyses of oxygen isotopes in 85 selected grains weremade using the CAMECA ims1270 ion microprobe atUCLA in March of 2009. A spot size of �20 lm and

Table 1In-day averages for 176Lu/177Hf and 176Hf/177Hf of the standard zirconsand Hergt (2005) shown for comparison. AS3 is called “FC-1” by Woodhfor the relevant quantity.

Quantity Accepted Session 1 2r Session 2 d. 1

Temora-2176Hf/177Hf 0.282686 0.282696 5.4e�5 0.282631176Lu/177Hf 0.00109 0.001059 8.1e�4 0.001055

Mud Tank176Hf/177Hf 0.282507 0.282486 2.1e�5 0.282487176Lu/177Hfa 0.000042 0.000049 2.4e�6 6.08e�6

AS3176Hf/177Hf 0.282184 0.282161 4.2e�5 0.282124176Lu/177Hf 0.001262 0.001178 5.8e�4 0.000931

a Not used for corrections.

primary beam current of ca. 1.5 nA were used (see Trailet al., 2007 for further analytical details). The AS3 zirconstandard (d18OSMOW = 5.34&; Trail et al., 2007) was usedfor correction of raw 18O/16O ratios. Oxygen isotope mea-surements made on grains for which multiple Hf–Pb do-mains were found are attributed to the Hf-Pb domainclosest to the surface of the grain, where the oxygen mea-surements were made. Data for all oxygen standards andunknowns are shown in EA-1.

3.4. Ti thermometry

Ti concentrations for Ti-in-zircon thermometry (Watsonand Harrison, 2005) on selected grains were made using theCAMECA ims1270 ion microprobe at UCLA in Marchand August of 2009. Details of the temperature-calculatingprocedure are given in Watson and Harrison (2005) andanalytical conditions for Ti measurement in Harrison andSchmitt (2007). We estimated uncertainties in our tempera-ture calculations by quadratic addition of the roughly 10 �Cuncertainty from the model of Watson and Harrison (2005)to uncertainty in our Ti concentrations. In calculatingapparent Txlln we assume that the TiO2 and SiO2 activityof the melt were 1 during zircon growth. The ubiquity ofquartz among mineral inclusions (see Section 4.1) providessupport for the latter assumption. Crystallization tempera-tures calculated for grains for which multiple Hf-age do-mains were found are attributed to the Hf-Pb domainclosest to the surface of the grain, where the Ti measure-ments were made.

4. RESULTS

4.1. Grain textures and zoning

The zircons display several zoning styles under cathodo-luminescence imaging. Whereas 57 grains show no zoning,32 show oscillatory zoning and 2 of the zircons show sectorzoning. Chaotic or patchy zoning is evident in 50 grains.Zircons with no zoning tend to be dark-to-medium in CLbrightness. Several 3.4–3.7 Ga grains have a well-definedrim-and-core geometry evident under CL. Obvious igneous(oscillatory or sector) zoning and pronounced rim/core

AS3, Temora-2 and Mud Tank. The reference values of Woodheadead and Hergt. Italics indicate the zircon was not used as a standard

2r Session 2 d. 2 2r Session 2 d. 3 2r

5e�6 0.282625 7.1e�5 0.282574 1.5e�46.8e�4 0.000932 4.5e�4 0.001039 6.4e�4

2.2e�5 0.282464 3.9e�5 0.282471 3.3e�51.0e�6 6.54e�6 1.3e�6 1.26e�5 1.5e�5

4.3e�5 0.282148 6.6e�5 0.282160 3.9e�51.9e�4 0.001292 8.5e�4 0.001136 4.2e�4

Concordant Zircons SHRIMP U-Pb Age N=3834 Holden et al. (2009)

Rel

ativ

e Pr

obab

ility

3.4 3.6 3.8 4.0 4.2 4.4

This studyLA-ICPMS Pb-Pb Age N=130

Rel

ativ

e Pr

obab

ility

207Pb/206Pb Age (Ga)

b

a

Fig. 6. Distributions in 207Pb/206Pb ages in the Jack Hills zirconsfrom (a) Holden et al.’s (2009) survey of Hadean Jack Hills zirconsusing SIMS U–Pb dating and (b) ages from this study. The small<4 Ga peaks (e.g., at 3.4 Ga) in the Holden et al. (2009) datarepresent zircons with initially Hadean-appearing 207Pb/206Pb agesthat upon closer analysis had younger cores.


geometry are rare in grains older than 3.6 Ga. Around 10%of zircons contain identifiable mineral inclusions at the sur-face, among which quartz and K-feldspar dominate.Quartz, K-feldspar, muscovite and biotite are present in3.4 Ga zircons while quartz is found in the 3.5–3.75 Ga zir-cons. One ilmenite inclusion was found in a 3.75 Ga zircon.All CL and inclusion information are given in Electronicannex EA-2.

4.2. Coupled Hf–Pb analyses

Results from the two analysis sessions agree well in bothLu–Hf systematics and ages despite their collection in dif-ferent laboratories (Fig. 4), highlighting the accuracy ofthe peak-stripping protocol. Nearly all of the 130 Hf–Pbdomain data points fall at realistic values of 176Lu/177Hffor zircon (<�0.001). Four analyses fall above 0.002 in176Lu/177Hf, which may indicate inaccuracies in the peak-stripping procedure for these few analyses (4/130; �3%).These analyses are marked in all figures. One of these sus-pect analyses has a 207Pb/206Pb age of �2.9 Ma and liesat ��28e. It is noticeably younger than the vast majorityof zircons found in previous studies of the Jack Hills; thisapparent age probably reflects ancient Pb loss. Its patchyzonation may also suggest alteration. Though its great dis-tance from DMM in eHf space may suggest a highly unrad-iogenic source, its high Lu/Hf renders further interpretationof the data point suspect. We do not include it in most ofour figures. Ages resemble the distribution shown in previ-ous studies (e.g., Crowley et al., 2005) with a large age peakca. 3.4 Ga and minor peaks ca. 3.45 and 3.55 Ga (Fig. 6).Fig. 7 displays the same data with the Hf–Pb domainsgrouped by the zonation style of their zircons.

We refer to the solar system initial 176Hf/177Hf ratio, be-low which no solar system samples should plot, as the Pri-mordial Hafnium Bound (PHB) for the rest of this workand compare our analyses to this bound. The proportionof domains displaying highly unradiogenic Hf (i.e., plottingnear PHB) decreases with decreasing age, such after 3.8 Gafew plot near PHB (see Fig. 7). Less than 4 Ga zircons with-in a few e units of PHB display textures under CL sugges-tive of a metamorphic origin, so the significance of theirunradiogenic hafnium in terms of igneous activity in thesource terrane(s) is uncertain. The distribution in207Pb/206Pb ages compares well to that observed in the restof the Jack Hills distribution, with a major peak near3.4 Ga and a minimum between 3.6 and 3.8 Ga (Holdenet al., 2009; Crowley et al., 2005; see Fig. 6). Hf-Pb datafor all samples can be seen in Electronic annex EA-2.

4.3. Oxygen isotopes

d18OSMOW values for the zircons average 5.49 ± 0.43&

(1r), in good agreement with the average mantle value of�5.3 ± 0.3& (1r, Valley, 2003). Ten zircons fall outsidethe range of common mantle values: seven zircons haved18O above 5.9& and three zircons are below 4.7&. Thereis no clear correlation of d18O with age, eHf or crystalliza-tion temperature (see Fig. 8). d18O values for all samplesare shown in EA-2.

4.4. Ti thermometry

Ti abundances in the zircons range from 0.89 to 36 ppm,indicating crystallization temperatures (Txlln) ranging from567 to 935 �C. Crystallization temperatures average679 ± 124 �C (2r). There is no apparent correlationbetween Txlln and age, eHf or d18O (see Fig. 8). Values of[Ti] and Txlln for all samples are shown in EA-2.

5. DISCUSSION

There are several notable differences in eHf, d18O, andTxlln between the Hadean zircon record in the Jack Hillsand the younger zircons sampled here. The age distributionin our zircons displays several discrete peaks rather than themore homogeneous age distribution seen in the Hadean (seeFig. 6). Zircons in the 3.4 Ga age peak exhibit a narrowrange in age with a wide range in eHf (–6 to –14). Minorage peaks at 3.45 and 3.5–3.6 Ga demonstrate similarbehavior. These groups behave similarly in d18O, exhibitinga range in d18O from 4.5 to 6.5& SMOW. There is no

3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1-20

-15

-10

-5

0

5

Ambiguous Rim/Core Metamorphic Igneous

207Pb/

206Pb Age (Ga)

Fig. 7. Jack Hills zircons from this study in eHf vs. age space, withzircons grouped by textures as imaged by cathodoluminescence.The “metamorphic” group consists of zircons with patchy ordisrupted zoning; the “igneous” group consists of zircons withoscillatory or sector zoning (or both). Oscillatory zonation is rare>3.6 Ga. Interestingly the most unradiogenic zircons <4 Ga displaymetamorphic textures, indicating that their hafnium compositionsprobably do not reflect source terrane magmatic evolution.

3

4

5

6

7

8

This study Cavosie et al. ('05) Peck et al. ('01) Trail et al. ('07) Harrison et al. ('08)

3.4 3.6 3.8 4.0 4.2

550

600

650

700

750

800

850

900

950

1000

1050 This study Harrison et al. (2008)

Txlln (C

)

207Pb/206Pb Age (Ga)

a

c

Fig. 8. Indicators for environment of formation among the studied Jack Hvs. Txlln. No correlations are apparent.


correlation between eHf and d18O within each of the agegroups. All groups exhibit large ranges in Txlln, indistin-guishable from the Hadean record.

Interestingly, our finding of almost exclusively mantle-like d18O among the <3.8 Ga zircons differs from the resultsfor similarly aged Jack Hills zircons reported by Peck et al.(2001). Peck et al.’s 3.3–3.6 Ga zircons average �6.3&

SMOW (n = 32 spot analyses on 16 grains). The discrep-ancy between these two datasets may owe partly to under-sampling in the earlier study – whereas Peck et al. analyzedonly 16 crystals with ages between 3.3 and 3.6 Ga for a totalof 32 analyses, we have analyzed 76 individual crystalsbetween the ages of 3.2 and 3.8 Ga, likely yielding a morerepresentative sample. It is also the case that we do notpre-screen our samples for U–Pb concordance, and givenTrail et al.’s (2007) and Booth et al.’s (2005) finding thatdiscordant Jack Hills zircons tend toward lower d18O thereis some danger that these grains may not reflect primarymagmatic d18O. However, the similarity of our averaged18O to that of 214 concordant Hadean Jack Hills zirconspreviously analyzed by ion microprobe (Peck et al., 2001;Cavosie et al., 2004; Trail et al., 2007; Harrison et al.,2008) suggests that any systematic error in the younger,non-screened zircons is likely not significant. Interestingly,several of our zircons fall below 5& SMOW (though within

4.4 -15 -10 -5 0 5 10

b

d

ills zircons. (a) age vs. d18O, (b) eHf vs. d18O, (c) age vs. Txlln, (d) eHf


the range of values found by previous studies), and mayindicate the remelting of higher temperature, hydrother-mally altered materials. However, it is not clear from thesmall sample size whether this is a significant part of theyounger zircon population, which mostly clusters aboutmantle values.

The variables of age, eHf and d18O can be used for prov-enance interpretations for the <4 Ga zircons. It is likelythat the large spreads in eHf at each age peak are due tothe mixing of material from multiple sources at differentinitial 176Hf/177Hf ratios to form the magma in question.This is consistent with the spread in d18O, which could re-flect minor source variations in d18O about a mantle value.The distinct clusters may suggest a low number of discretesource units. The ages of the younger clusters are consistentwith the ages of some orthogneiss units in the NarryerGneiss Complex (Kinny et al., 1988; Myers, 1988a,b; Nut-man et al., 1991). Zircons from the Dugel Gneiss reveal acrystallization age of 3375 ± 26 Ma, the Eurada Gneissyields zircon ages of �3.45–3.49 Ga, and zircons from thevarious portions of the Meeberrie Gneiss range from youn-ger than 3.40 to �3.73 Ga (Nutman et al., 1991). In com-parison to the <3.6 Ga zircons, the Hadean record wouldbe more suggestive of derivation from a multitude ofsources. This smoother distribution in the Hadean detritalzircons could be effected also if the Hadean zircons haveundergone multiple sedimentary cycles – the populationthus likely deriving from a larger total area and numberof protoliths.

Despite these consistent ages, others (e.g., Maas andMcCulloch, 1992) have pointed to geochemical discrepan-cies between the detrital zircon-bearing metasedimentsand the Narryer orthogneisses that may indicate the gneis-ses are not the source of the younger detrital zircons. TheNarryer orthogneisses display HREE depletion and widelyvariant positive and negative Eu anomalies, in contrast to

a

Fig. 9. Jack Hills zircons in age vs. eHf space from this and several previostudies. The “PHB” line represents the evolution of a reservoir with the sthigh Lu/Hf analysis at 2.88 Ga, �28e is omitted. (b) Our 3.2 – data sComplex. Data from several of these studies were normalized to slightly diBouvier et al. (2008) for a more apt comparison to our data.

the unfractionated HREE and consistently negative Euanomaly in the metasediments. However, these differencescould be balanced by a high proportion of mafic sediment(also suggested by detrital chromite and high Cr, Ni inthe metasediments). A more relevant dataset for compari-son was compiled by Kemp et al. (2010), who report ageand Lu–Hf information for 70 metaigneous zirconsseparated from gneisses in the Narryer Gneiss Complex.Fig. 9 shows our data in the context of both (a) earlier stud-ies of detrital Jack Hills zircons and (b) the metaigneous zir-cons of Kemp et al. (2010). Narryer orthogneiss zirconsrange in age from 2.6 to 3.75 Ga and overlap with the moreradiogenic of the <3.8 Ga Jack Hills zircons, though thereare slight discrepancies in the peak ages. It appears that alarge proportion of the more radiogenic <4 Ga Jack Hillszircons are at least consistent with derivation from localgneisses. The lack of more unradiogenic zircons amongthe Narryer gneisses would prove more puzzling in this con-text, perhaps requiring a less radiogenic reservoir of mate-rial, no longer outcropping in the area, which recordedmany of the same events 3.3–3.75 Ga as the Narryer gneis-ses. It may be that the metasediments’ derivation from thiswider set of source lithologies accounts for the chemical dis-crepancies observed by Maas and McCulloch (1992),though the large mafic component needed to balance theREE discrepancies between Narryer gneisses and themetasediments may not be an excellent source for the largecomponent of radiogenic zircons. However, due to the stillunconstrained nature of the relationship between the JackHills source terrane(s) and the extant Narryer gneisses,for this study we consider the identity of the younger zir-cons’ protoliths to remain an open question. For theremainder of this paper we will draw conclusions aboutthe geochemistry and crustal evolution of the source regionfor the Jack Hills metaconglomerate, whatever thissource(s) might be in terms of extant rock units.

b3.2 3.3 3.4 3.5 3.6 3.7 3.8

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

us studies. (a) Our data compared to other Jack Hills detrital zirconated 176Lu/177Hf separated from CHUR at 4560 Ga. As in Fig. 4, ahown with metaigneous zircon analyses from the Narryer Gneissfferent CHUR values; we have renormalized to the CHUR values of


5.1. Implications of initial hafnium isotopes and model

extraction ages

Model mantle extraction ages for the crust representedin any zircon population are dependent upon the choiceof 176Lu/177Hf ratio for the crust. In the case of detrital zir-con populations, the original host rocks are not availablefor analysis and reasonable assumptions must be maderegarding 176Lu/177Hf for the grains in question. The sim-plest method involves assuming a uniform 176Lu/177Hf forthe population in the absence of contrary evidence. We con-sider several models with differing values (Fig. 10a–c). Thefirst model (Fig. 10a) uses 176Lu/177Hf = 0.01, which isclose to the average for volcanic rocks from the GEOROCdatabase (http://georoc.mpch-mainz.gwdg.de/georoc/Star-t.asp; see Sarbas, 2008) and consistent with the bulk crustvalue of 0.008 preferred by Rudnick and Gao (2003). A sec-ond model (Fig. 10b) explores the possibility of a mafic ori-gin for the zircons with 176Lu/177Hf = 0.022. A third model

Fig. 10. Several models for zircon Lu–Hf extraction ages. (a) Uniform mUniform 176Lu/177Hf = 0.006. (d) A basis for more complex models incorand mafic reservoirs are shown originating at 4.56 Ga, along with a n176Lu/177Hf = 0.006 reservoir fit to the <3.6 Ga distribution. Analyses wshould be regarded with caution; one at 2.88 Ga and �28e is omitted. ThFig. 9a from Amelin et al. (1999), Harrison et al. (2005, 2008), Blichert-

(Fig. 10c) uses a value derived by fitting a reservoir evolu-tion line to the least radiogenic of the <3.8 Ga zircons withigneous zonation, which yields a value of �0.006. Depletedmantle parameters were calculated by extrapolating froman assumed present eHf value of +18 to a value of zero at4.56 Ga, assuming no appreciable external changes to the176Lu/177Hf of the reservoir apart from the decay of 176Luover geologic time. As an additional note of caution, inthe case of source mixing during magma formation, as islikely for igneous zircons at 3.4 Ga, the model age forany particular zircon may not have much significance andmodel ages for zircons from the extreme ends of the eHf dis-tribution will be most significant in terms of source materialextraction age.

5.1.1. Uniform models

The two more felsic models with 176Lu/177Hf = 0.01(average felsic volcanics) and 176Lu/177Hf = 0.006 (best fitto younger distribution) generally yield depleted mantle

odel with 176Lu/177Hf = 0.01. (b) Uniform 176Lu/177Hf = 0.022. (c)porating multiple past reservoirs. See the text in Section 5.1.2. Felsicecessary juvenile reservoir sometime <4.2 Ga and a hypothetical

ith unusually high 176Lu/177Hf are marked with gray hexagons ande data marked “earlier studies” are the detrital zircon data shown inToft and Albarede (2008), and Kemp et al. (2010).

http://georoc.mpch-mainz.gwdg.de/georoc/Start.asp

http://georoc.mpch-mainz.gwdg.de/georoc/Start.asp


extraction ages (TDM’s) for the <3.6 Ga materials between3.8 and 4.3 Ga. At �3.8 Ga there is a fairly abrupt transi-tion in both models in the average TDM. Formation of all3.4–3.45 Ga zircons directly from a long-lived mafic reser-voir of 176Lu/177Hf = 0.022 (a value derived by Amelinet al., 1999 from the slope of their age, eHf array) is unlikelygiven the unrealistically high model ages of >4.56 Ga forthe most unradiogenic zircons. The most unradiogenic3.4 Ga zircons are also (as shown in Fig. 7) magmaticallyzoned, demonstrating that they represent igneous materialsrather than older material metamorphosed at 3.4 Ga, whichis a likely origin for the patchily zoned <4 Ga zircons plot-ting near PHB. A direct mafic source is also inconsistentwith the observed granitic mineral inclusions (quartz + K-feldspar + micas) and with minimum melting conditions in-ferred from low Txlln among the zircons.

5.1.2. More complex extraction age models

One possibility for the derivation of the younger zirconsis a granitic source only recently remelted from an older,long-lived mafic source. In the above calculations andFig. 10a–c we assume that each modeled reservoir was ex-tracted directly from the mantle. This is a simplified model,as direct mantle melts would have typical basaltic Lu/Hf ra-tios (higher than felsic, but with variations depending ongarnet content). The two lower Lu/Hf reservoirs wouldmost likely have formed from a mafic reservoir that was ex-tracted from the mantle at some unspecified time in the past– and thus the calculated extraction ages for the two lowerLu/Hf reservoirs should be viewed as minimum model ages.This does not, however, erase the necessity for the mostunradiogenic Hadean zircons and the far more radiogenic<3.6 Ga zircons to be derived from distinct sourcereservoirs.

Fig. 10d shows the evolution of hypothetical reservoirsthat might form the basis for a more complex model of JackHills source terrane(s) development, including reservoirswith the same 176Lu/177Hf ratios investigated in the uni-form-composition models. Felsic and mafic reservoirs iso-lated at planet formation permissively bracket the moreunradiogenic among the sampled Jack Hills materials,though as shown in Fig. 10d no <3.6 Ga magmatic grainsapproach the felsic evolution line. Mafic reservoirs ex-tracted at 4.56 and 4.0 Ga bracket most 3.5–4.0 Ga zirconsand the more unradiogenic 3.4–3.5 Ga zircons, thoughsome felsic history is required for zircons below ��10e at3.45 and 3.4 Ga (many of which display igneous zonation).A variety of possible source terrane evolutions are shown.For example, a hypothetical history involving the extrac-tion of a felsic reservoir from a 4.56 Ga mafic reservoir isshown to be consistent with the <3.6 Ga data; so alsowould be mixtures (of varying degrees) of ancient maficand felsic reservoirs.

Despite the many working hypotheses consistent withthese data, several things are clear from this figure. First,the <3.6 Ga materials must be derived largely from differentsources than the unradiogenic Hadean zircons – eitheryounger materials with a long felsic history or felsic materi-als derived more recently from a very ancient mafic reser-voir. Fig. 10d shows the latter scenario for the derivation

of unradiogenic 3.4 Ga magmas. The most unradiogenicHadean material is not unambiguously sampled after3.6 Ga. At the same time, model ages show that the moreunradiogenic zircons (below �10e at 3.4 Ga) must havebeen sourced from a reservoir with Lu/Hf lower than typi-cal mafic rocks, though not inconsistent with ancient re-melts of mafic crust. Second, unlike most of the sampledHadean zircons (Harrison et al., 2005, 2008; Kemp et al.,2010), the more radiogenic materials among the <3.8 Gadistribution require juvenile mantle melts at some point inthe late Hadean to post-Hadean history of the source ter-rane (with the present zircons sourced partially from rewor-kings of these mantle melts). Using Fig. 9, estimates of thetiming for the simplest model – a direct mafic source – fall�4 Ga.

5.2. Formation environment and petrogenesis

The relatively narrow distribution of the zircons about amantle-like d18O value (5.49 ± 0.43& d18OSMOW, ±1r)suggests little systematic contribution to the host rocksfrom materials involved in low-temperature aqueous alter-ation. For the igneous zircons, this likely precludes a largedegree of (meta-) sediment assimilation into host magmas.These results are in contrast to the much more variantHadean oxygen record which despite the similar averaged18OSMOW value of 5.72& has a higher degree of variationat ±0.8& (±1r). The Hadean record contains many grainsfalling above 6.5& that likely reflect significant contribu-tions from older sedimentary materials (see Fig. 8a).

The crystallization temperatures recorded in the zirconTi abundances of 679 ± 62 �C (1r) suggest that, like themajority of Hadean zircons, the younger population if igne-ous represent close to minimum melting conditions of inter-mediate to felsic magmas. An igneous origin is indeedsuggested by the oscillatory zoning evident in 32 grains.The 107 grains displaying patchy, irregular, or no zoningare more ambiguous in origin. The 50 patchily-zoned grainsmay have undergone recrystallization or metamorphicoverprinting (Corfu et al., 2003), which might obscure theoriginal igneous Hf and O isotopic compositions, but nev-ertheless except in the noted cases of highly unradiogenicpatchy grains these apparently metamorphic grains do notdiffer systematically from other zircons in their geochemis-try. Though there is no apparent systematic relationshipamong the distributions of d18O or eHf and zoning style,grains with Txlln <600 �C tend to display either no zonationor patchy/irregular zonation, though these zonation catego-ries also include higher-Txlln grains. From 3.4 to 3.6 Ga,patchily and irregularly zoned zircons occur in the sameage intervals as the apparently more pristine oscillatorilyzoned zircons, and zircons with no zoning occur alongsidethem.

5.3. Global comparison

Pietranik et al. (2008) compiled a Lu–Hf-age dataset ofzircons from several Archean cratons and interpret it toreveal several pulses of continental growth between 4.5and 2.8 Ga. Many zircons in this compilation are


contemporaneous with the younger Jack Hills zircons andmay provide a comparison for other regions where earlyArchean crustal evolution left some lithic record. We plotour age vs. eHf data with Pietranik et al.’s (2008) data forSlave Craton detrital zircons and Amelin et al.’s (2000) datafor detrital zircons from the Acasta Gneiss, BarbertonMountain Land, Pilbara Craton and Itsaq Gneiss(Fig. 11). We have normalized all hafnium compositionsto the CHUR values of Bouvier et al. (2008) for compari-son. Zircons with eHf values as low as –10 occur in theAcasta Gneiss from 3.6 to 3.4 Ga. Interestingly, the sam-pled Acasta Gneiss zircons overlap considerably in eHf withzircons from our 3.55 Ga broad age peak and may indicateparallel crustal reworking histories in the two terranes forthis time period.

Pietranik et al. (2008) interpret Mid- to Late-Archeanzircons from the Slave Craton to lie along a mafic reservoirevolution line that suggests a mantle extraction age of�4.2 Ga. Many of our samples are also compatible withorigins from a remelted mafic Hadean reservoir and overlapthe reported compositions of Slave and Superior provincezircons. However, our 3.4 Ga zircons reach much moreunradiogenic compositions than those found in other cra-tons for the same time period. It is likely that the unusuallyunradiogenic younger Jack Hills zircons are a unique re-source in determining the fate of early felsic crust. This iscompounded by the apparent lack of highly unradiogenicsignatures among detrital zircons of similar age from thenearby Mt. Narryer (Nebel-Jacobsen et al., 2010).

5.4. Synthesis: crustal evolution from 3.0 to 4.0 Ga

While a powerful approach for elucidating juvenile crus-tal addition vs. remelting of older crust, Lu–Hf isotopic

Fig. 11. Jack Hills detrital zircons from this and previous studiescompared with zircons of similar age from other Archean cratons.Slave Craton data are from Pietranik et al. (2008); Acasta Gneiss,Barberton Mountain Land, and Pilbara Craton zircons are fromAmelin et al. (2000); Mt. Narryer detrital zircon data are fromNebel-Jacobsen et al. (2010); earlier Jack Hills detrital zircon dataas for Fig. 10d. We normalized all Lu–Hf data to the CHURparameters of Bouvier et al. (2008) for comparison to our data.

data do not constrain the tectonic environment in whichthe metaconglomerate catchment area evolved. The appar-ent contrasts between older and younger zircons in bothLu–Hf and oxygen isotope systematics may suggest differ-ent magmatic settings and possibly different reservoirs ofmaterial for <3.6 and >4 Ga protoliths. Further study of<4 Ga Jack Hills zircons and the Narryer Gneiss Complexmay be needed to constrain the specific geologic context inwhich the Hadean grains have been preserved and theyounger grains formed, but some constraints are evidentnow.

The change in Lu–Hf systematics with age precludes thepreservation or remelting of significant amounts of themost enriched sampled Hadean materials (that falling onor near the PHB) into younger magmas and requires atleast some contribution from late- to post-Hadean juvenilecrust. Why this unradiogenic reservoir ceased to be avail-able for reworking is unclear. Solutions involving the ero-sion and loss of this crustal material (for instance, bysome form of tectonic denudation) are intriguing but awaitfurther analysis of the period 3.6–4 Ga to determinewhether any magmatic zircons bearing this signature areevident in a more representative sampling. This period issparsely sampled at present due to its coincidence with aminimum in the zircon age distribution at 3.6–3.8 Ga.

Despite the transition in the Lu–Hf systematics withtime, a large separation in model ages for Hadean andyounger zircons is only achieved with a felsic precursorfor the younger magmas. The isotopic evidence does notrule out an alternate scenario in which most <3.6 Ga zir-cons derive from a remelted mafic Hadean precursor indis-tinguishable from the more radiogenic of the sampledHadean materials. Derivation from remelting of maficHadean materials is also consistent with the Lu–Hf system-atics of contemporaneous zircons from several spatiallydistinct cratons (Amelin et al., 2000; Pietranik et al.,2008; Fig. 11), and is likely a common petrogenesis amongEarly Archean crust. We must also consider the possibilityof a mixture of various sources, Hadean and younger, in theproduction of younger magmas. Further sampling 3.6–4 Gawill constrain the form of the distribution and determinewhether the younger Jack Hills zircons are indeed bestmodeled by magmas derived from felsic or mafic precursormaterials (or a mixture).

Whether the melts from which the <3.6 Ga Jack Hillszircons formed are remelts of mafic or felsic materials theyappear to be largely felsic themselves. Ti-in-zircon Txlln

throughout the age distribution suggest close to minimummelting conditions. There is no obvious distinction betweenthe Hadean and younger populations in Txlln. The mineralinclusion suite is only well characterized for the 3.4 Ga agepeak, but here appears largely consistent with a graniticmelt: quartz and K-feldspar dominate. The largely man-tle-like d18O among <3.6 Ga zircons suggest little contribu-tion of sedimentary material to the magmas. This iscorroborated for grains in the 3.4 Ga age peak by the smal-ler proportion of muscovite among mineral inclusions com-pared to the Hadean (e.g., Hopkins et al., 2008, 2010). Theapparently granitic Txlln combined with the mantle-liked18O of most grains suggests the <3.6 Ga zircons are likely


derived from I-type granitic melts. In contrast, a significantproportion of the Hadean grains appear on the basis ofmineral inclusions to be from S-type melts (Hopkinset al., 2008, 2010). This suggests a transition to a differentmagmatic style with possible tectonic implications. Boththe Lu–Hf systematics and other geochemical indicatorsof petrogenesis independently suggest a transition betweenthe Hadean and <3.6 Ga source areas. Whether this isdue to the zircons deriving from separate tectonic terranesor one region undergoing continuous geologic evolutionaway from S-type granite production is yet unclear.

6. CONCLUSIONS

Age, eHf, and d18O results for <3.8 Ga zircons in the JackHills detrital record reveal important differences between theHadean and younger zircon populations. Whereas themajority of igneous zircons from both time periods appearto result from minimum melting of a low Lu/Hf source, the<3.6 Ga zircons lack the highly unradiogenic materials sam-pled in the Hadean, require at least some juvenile input notseen among the Hadean zircons, and appear less likely to re-flect aqueous alteration (hydrothermal or low-T) amongtheir source materials. This indicates a substantial prove-nance difference between the Hadean and younger grains.Whether this indicates separate tectonic terranes remains tobe seen. For petrogenesis of the <3.6 Ga magmas by remelt-ing of an older felsic source, there is an abrupt transition ofthe eHf distribution at 3.6–3.8 Ga; models based on theremelting of Hadean mafic materials for some of the youngerzircons remove the abrupt transition but still require disap-pearance of the most unradiogenic Hadean materials. A lar-ger dataset for the period from 3.6 to 4 Ga will clarify the fateof the unradiogenic Hadean component and further con-strain possible tectonic scenarios accounting for the eHf dis-tribution in the younger Jack Hills record. Further surveysof geochemistry among the older and younger zircon popu-lations will also help to elucidate the relationship betweenthe two groups. The 3.6–4 Ga portion of the Jack Hills recordis potentially one of the best resources on the planet for prob-ing this poorly understood period of early Earth history,especially given that <3.6 Ga zircons preserve some of themost unradiogenic hafnium seen in the early Archean. Fur-ther constraining the geologic behavior of the source region3.6–4 Ga will have important implications for the tectonicevolution of this fragment of the very early continental crust.

ACKNOWLEDGMENTS

We thank Dianne Taylor, Karen Ziegler, and Axel Schmitt forinvaluable instruction in analyzing our samples. This research wasconducted with support from a grant from NSF-EAR’s Petrology/Geochemistry Program to T.M.H. and an NSF Graduate ResearchFellowship (DGE-0707424) to E.A.B.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.gca.2011.06.007.

REFERENCES

Amelin Y. V. (1998) Geochronology of the Jack Hills detritalzircons by precise U–Pb isotope dilution analysis of crystalfragments. Chem. Geol. 146, 25–38.

Amelin Y. V. (2005) Meteorite phosphates show constant Lu-176decay rate since 4557 million years ago. Science 310, 839–841.

Amelin Y. V., Lee D. C. and Halliday A. N. (2000) Early-MiddleArchean crustal evolution deduced from Lu–Hf and U–Pbisotopic studies of single zircon grains. Geochim. Cosmochim.

Acta 64, 4205–4225.

Amelin Y. V., Lee D. C., Halliday A. N. and Pidgeon R. T. (1999)Nature of the Earth’s earliest crust from hafnium isotopes insingle detrital zircons. Nature 399, 252–255.

Bennett V. C., McCulloch M. T., Nutman A. P. and Kinny P. D.(1990) Combined Sm–Nd isotopic systematics and U–Pb zirconcharacteristics in the Narryer Gneiss complex, Western Aus-tralia: implications for Archean crustal genesis. Intl. Archean

Symp. 3, 9–10.

Blichert-Toft J. and Albarede F. (2008) Hafnium in Jack Hillszircons and the formation of the Hadean crust. Earth Planet.

Sci. Lett. 265, 686–702.

Booth A. L., Kolodny Y., Chamberlain C. P., McWilliams M.,Schmitt A. K. and Wooden J. (2005) Oxygen isotopic compo-sition and U–Pb discordance in zircon. Geochim. Cosmochim.

Acta 69, 4895–4905.

Bouvier A., Vervoort J. D. and Patchett P. J. (2008) The Lu–Hfand Sm–Nd isotopic composition of CHUR: constraints fromunequilibrated chondrites and implications for the bulk com-position of terrestrial planets. Earth Planet. Sci. Lett. 273, 48–

57.

Boyet M. and Carlson R. W. (2005) 142Nd evidence for early (>4.53billion years ago) global differentiation of the silicate Earth.Science 309, 576–581.

Cates N. L. and Mojzsis S. J. (2009) Metamorphic zircon, traceelements, and Neoarchean metamorphism in the ca. 3.75 GaNuvvuagittuq supracrustal belt, Quebec (Canada). Chem. Geol.

261, 98–113.

Caro G., Bourdon B., Birck J. and Moorbath S. (2003) Sm–Ndevidence from Isua metamorphosed sediments for early differ-entiation of the Earth’s mantle. Nature 423, 428–432.

Cavosie A. J., Wilde S. A., Liu D., Weiblen P. W. and Valley J. W.(2004) Internal zoning and U–Th–Pb chemistry of Jack Hillsdetrital zircons: a mineral record of early Archean to Meso-proterozoic (4348–1576 Ma) magmatism. Precambr. Res. 135,

251–279.

Compston W. and Pidgeon R. T. (1986) Jack Hills, evidence ofmore very old detrital zircons in Western Australia. Nature 321,

766–769.

Corfu F., Hanchar J. M., Hoskin P. W. O. and Kinny P. (2003)Atlas of zircon textures. In Zircon. Rev. Mineral. (eds. J. M.Hanchar and P. W. O. Hoskin) 53, 469–500.

Crowley J. L., Myers J. S., Sylvester P. J. and Cox R. A. (2005)Detrital zircon from the Jack Hills and Mount Narryer,Western Australia: evidence for diverse >4.0 Ga source rocks.J. Geol. 113, 239–263.

Dauphas N. and Chaussidon M. (2011) A perspective from extinctradionuclides on a young stellar object: the sun and its accretiondisk. Ann. Revs. Earth Planet. Sci. 39, 351–386.

DeLaeter J. L., Fletcher I. R., Bickle M. J., Myers J. S., Libby W.G. and Williams I. R. (1985) Rb–Sr, Sm–Nd and Pb–Pbgeochronology of ancient gneisses from Mt. Narryer, WesternAustralia. Aust. J. Earth Sci. 32, 349–358.

Finch R. J. and Hanchar J. M. (2003) Structure and chemistry ofzircon and zircon-group minerals. In Zircon. Rev. Mineral. (eds.J. M. Hanchar and P. W. O. Hoskin) 53, 1–25.




Fletcher I. R., Rosman K. J. R. and Libby W. G. (1988) Sm–Nd,Pb–Pb and Rb–Sr geochronology of the Manred Complex,Mount Narryer, Western Australia. Precambr. Res. 38, 343–

354.

Harrison T. M., Blichert-Toft J., Muller W., Albarede F., HoldenP. and Mojzsis S. J. (2005) Heterogeneous Hadean hafnium:evidence of continental crust by 4.4–4.5 Ga. Science 310, 1947–

1950.

Harrison T. M. and Schmitt A. K. (2007) High sensitivity mappingof Ti distributions in Hadean zircons. Earth Planet. Sci. Lett.

261, 9–19.

Harrison T. M., Schmitt A. K., McCulloch M. T. and Lovera O.M. (2008) Early (P4.5 Ga) formation of terrestrial crust: Lu–Hf, d18O, and Ti thermometry results for Hadean zircons. Earth

Planet. Sci. Lett. 268, 476–486.

Holden P., Lanc P., Ireland T. R., Harrison T. M., Foster J. J. andBruce Z. (2009) Mass-spectrometric mining of Hadean zirconsby automated SHRIMP multi-collector and single-collector U/Pb zircon age dating: the first 100,000 grains. Int. J. Mass Spec.

286, 53–63.

Hopkins M., Harrison T. M. and Manning C. E. (2008) Low heatflow inferred from >4 Gyr zircons suggests Hadean plateboundary interactions. Nature 456, 493–496.

Hopkins M., Harrison T. M. and Manning C. E. (2010)Constraints on Hadean geodynamics from mineral inclusionsin >4 Ga zircons. Earth Planet. Sci. Lett. 298, 367–376.

Kemp A. I. S., Wilde S. A., Hawkesworth C. J., Coath C. D.,Nemchin A., Pidgeon R. T., Vervoort J. D. and DuFrane S. A.(2010) Hadean crustal evolution revisited: new constraints fromPb–Hf isotope systematics of the Jack Hills zircons. Earth


Kinny P. D., Williams I. S., Froude D. O., Ireland T. R. andCompston W. (1988) Early Archaean zircon ages from orthog-neisses and anorthosites at Mount Narryer, Western Australia.Precambr. Res. 38, 325–341.

Kober B., Pidgeon R. T. and Lippolt H. J. (1989) Single-zircondating by stepwise Pb-evaporation constrains the Archeanhistory of detrital zircons from the Jack Hills, WesternAustralia. Earth Planet. Sci. Lett. 91, 286–296.

Maas R. and McCulloch M. T. (1992) The provenance of Archeanclastic metasediments in the Narryer Gneiss Complex, WesternAustralia: trace element geochemistry, Nd isotopes, and U–Pbages for detrital zircons. Geochim. Cosmochim. Acta 55, 1915–

1932.

Mojzsis S. J., Harrison T. M. and Pidgeon R. T. (2001) Oxygen-isotope evidence from ancient zircons for liquid water at theEarth’s surface 4300 Myr ago. Nature 409, 178–181.

Myers J. S. (1988a) Early Archaean Narryer Gneiss complex,Yilgarn Craton, Western Australia. Precambr. Res. 38, 297–

307.

Myers J. S. (1988b) Oldest known terrestrial anorthosite at MountNarryer, Western Australia. Precambr. Res. 38, 309–323.

Nebel-Jacobsen Y., Munker C., Nebel O., Gerdes A., Mezger K.and Nelson D. R. (2010) Reworking of Earth’s first crust:Constraints from Hf isotopes in Archean zircons from Mt.Narryer, Australia. Precambr. Res. 182, 175–186.

Nutman A. P., Kinny P. D., Compston W. and Williams I. S.(1991) SHRIMP U–Pb zircon geochronology of the NarryerGneiss Complex, Western Australia. Precambr. Res. 52, 275–

300.

O’Neill J., Carlson R. W., Francis D. and Stevenson R. K. (2008)Neodymium-142 evidence for Hadean mafic crust. Science 321,

1828–1831.

Peck W. H., Valley J. W., Wilde S. A. and Graham C. M. (2001)Oxygen isotope ratios and rare earth elements in 3.3–4.4 Gazircons: ion microprobe evidence for high d18O continentalcrust and oceans in the Early Archean. Geochim. Cosmochim.

Acta 65, 4215–4299.

Pietranik A. B., Hawkesworth C. J., Storey C. D., Kemp A. I. S.,Sircombe K. N., Whitehouse M. J. and Bleeker W. (2008)Episodic, mafic crust formation from 4.5 to 2.8 Ga: newevidence from detrital zircons, Slave craton, Canada. Geology

36, 875–878.

Rudnick R. L. and Gao S. (2003) Composition of the continentalcrust. In The Crust, Treatise on Geochemistry (ed. R.L.Rudnick) 3, 1–64.

Sarbas B. (2008) The GEOROC database as part of a growinggeoinformatics network. In Geoinformatics 2008 – Data to

Knowledge. Proceedings: U.S. Geological Survey Scientific

Investigations Report (eds. S. R. Brady, A. K. Sinha and L.C. Gundersen) 2008-5172, 42–43.

Scherer E., Mezger K. and Munker C. (2004) The 176Lu decayconstant revisited. Geochim. Cosmochim. Acta 68(Suppl.),

A748.

Soderlund U., Patchett J. P., Vervoort J. D. and Isachsen C. E.(2004) The 176Lu decay constant determined by Lu–Hf and U–Pb isotope systematics of Precambrian mafic intrusions. Earth


Spaggiari C. V., Pidgeon R. T. and Wilde S. A. (2007) The JackHills greenstone belt, Western Australia Part 2: lithologicalrelationships and implications for the deposition of P4.0 Gadetrital zircons. Precambr. Res. 155, 261–286.

Taylor D. J., McKeegan K. D. and Harrison T. M. (2008)176Lu–177Hf zircon evidence for rapid lunar differentiation.Earth Planet. Sci. Lett. 279, 157–164.

Taylor S. R. and McLennan S. M. (1985) The Continental Crust:

Its Composition and Evolution. Blackwell, Oxford, UK, 311 p..Thirlwall M. F. and Anczkiewicz R. (2004) Multidynamic isotope

ratio analysis using MC-ICP-MS and the causes of secular driftin Hf, Nd and Pb isotope ratios. Int. J. Mass Spec. 235, 59–81.

Trail D., Mojzsis S. J., Harrison T. M., Schmitt A. K., Watson E.B. and Young E. D. (2007) Constraints on Hadean zirconprotoliths from oxygen isotopes, REEs and Ti-thermometry.Geochem. Geophys. Geosyst. 8, Q06014.

Valley J. W. (2003) Oxygen isotopes in zircon. In Zircon. Rev.

Mineral. (eds. J.M. Hanchar and P.W.O. Hoskin) 53, 343–385.Watson E. B. and Harrison T. M. (2005) Zircon thermometer

reveals minimum melting conditions on earliest Earth. Science

308, 841–844.

Williams I. R. and Myers J. S. (1987) Archean geology of theMount Narryer region, Western Australia. W. Aust. Geol. Surv.

Rep. 22, 32.

Woodhead J. D. and Hergt J. M. (2005) A preliminary appraisal ofseven natural zircon reference materials for in situ Hf isotopedetermination. Geostd. Geoanal. Res. 29, 183–195.

Woodhead J. D., Hergt J. M., Shelley M., Eggins S. and Kemp R.(2004) Zircon Hf isotope analysis with an excimer laser, depthprofiling, ablation of complex geometries, and concomitant ageestimation. Chem. Geol. 209, 121–135.

Xie L., Zhang Y., Zhang H., Sun J. and Wu F. (2008) In situsimultaneous determination of trace elements, U–Pb and Lu–Hf isotopes in zircon and baddeleyite. Chin. Sci. Bull. 53, 1565–

1573.

Associate editor: F.J. Ryerson

bell e., harrison t.m., mcculloch m.t. and young e.d. (2011)

Documents